Making sounds with analogue electronics - Part 9: Sampling in an analogue environment

'Bucket-brigade' delay lines
The most common technology which met these requirements in the 1970s was the 'bucket-brigade' delay line or analogue delay line. This used the charge on a series of capacitors to represent the audio signal, rather than the magnetic field used in tape systems or the numbers used in digital systems.

The sampling process was merely the opening of an electronic switch to charge up the first capacitor in the delay line. The size of the voltage determined the amount of charge that was transferred to the capacitor: the higher the voltage which was being sampled, the more the charge which was stored in the capacitor. Effectively, the capacitor acted as a store for the voltage, since the presence of the charge in the capacitor was shown by the voltage across the capacitor. The switch then opened and the charge was held in the capacitor since there was no significant leakage path.

Another switch was then used to transfer the charge to the next capacitor in the delay line, where it again produced a voltage. The original capacitor was then available to sample the next point on the incoming audio signal. This process continued, with the sample voltages moving along the delay line formed by the capacitors; hence the term 'bucket-brigade' delay lines (Figure 3.8.3).

Because each section of the delay line is just a capacitor and some electronic switches, it was easy to fabricate, and so several thousands could be placed on a single IC chip. The sampling and transfer of charges required a relatively high-frequency clock signal, but the control circuitry was straightforward. This simplicity of control and application made analogue 'bucket-brigade' delay lines popular in the 1970s and early 1980s for producing echo, chorus and reverberation effects. At least one monophonic sampler was produced using analogue delay lines in the early 1980s, but it was rapidly superseded by digital versions.

The limitations of the analogue delay line technique are many fold: first, the capacitors and switches are not perfect, so some of the charge leaks away causing signal loss, distortion and noise; but more importantly, the high-frequency clock signals tend to become superimposed on the output audio signals and this degrades the usable dynamic range of the delay line.

Also, because they sample the audio, the high-frequency sample clock needs to be low in order to achieve long time delays, but then the clock rate interferes with the audio signal. At high clock rates, the delay time is short. And so they acquired a reputation for poor high-frequency response, which was a direct result of designs that sampled at too low a frequency in order to try and maximize the delay time. Because of these problems, digital sampling technology has replaced analogue delay lines and modern equivalents can easily put an (analogue-to-digital converter ADC), (digital-to-analogue converter DAC) and storage onto a single chip.

As with many synthesizer-related analogue chips, some bucket-brigade delay line chips are now rare and can sometimes attract high prices when they are needed to repair old guitar flanger/chorus/echo units.

Delay lines
An alternative to bucket-brigade delay lines moving charge around is to use metal springs or metal plates to carry the sound signals acoustically/mechanically. Sounds are transferred to the metal using modified loudspeaker drivers, and the delayed sound signals are recovered with contact microphones.

The physical size of these acoustic delay lines can be large, and the 'spring lines' and 'plate echoes' of the 1960s and 1970s have again been largely replaced by digital alternatives, including many emulations! Acoustic delay lines have the advantage that they are not a sampling system, but are more suited to reverberation effects than pure sampling – they are not suited to storing a sound and subsequently replaying it, instead they simply store a sound for short time.

Optical
One alternative sampling method uses a technique which is similar in principle to tape recording. Optical film soundtracks are a light-based variation of tape recording. Instead of storing the audio as a changing magnetic field, the film soundtrack uses the amount of light passing through the film to store the audio signal. This is normally achieved by arranging for large audio signal levels to allow a large amount of light to pass through the film, whilst small signals allow less light through. A photodetector and lamp are used to convert the transmission of light into an audio signal.

This modulation of light by an audio signal is normally achieved by using the audio waveform to control the width of a slot, and so the amount of light that passes through the film. Variable density (opacity) film can also be used, but this is rare for film use, although it has been used for experimental systems where film is used to produce sound by literally painting onto it to control the amount of light that passes through it at any given instant. By passing the resulting film through a lamp and photodetector, the optical version of the audio can be converted into sound. Although flexible, the complexity of producing the required degree of detail is enormous and very time consuming.

At least one manufacturer produced an optical sample-replay machine in the 1980s, but as with all analogue methods, this was not a success against the digital competitors.

Coming up in Part 10: Sequencing, wiring, recording, performing, example instruments and timeline.

Printed with permission from Focal Press, a division of Elsevier. Copyright 2009. "Sound Synthesis and Sampling" by Martin Russ. For more information about this title and other similar books, please visit www.elsevierdirect.com.

For more articles like this and others related to audio design, visit Audio Designline and/or subscribe to the monthly Audio newsletter (free registration).